Single Shot MultiBox Detector for Urban Plantation Single Tree Detection and Location With High-Resolution Remote Sensing Imagery

Zheng, Yueyuan; Wu, Gang

doi:10.3389/fenvs.2021.755587

Cited by 8 publications

(5 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To develop a tracking method that is simple, and robust to illumination changes and long-term occlusions, machine learning approaches are used to detect animals or objects of interest in video frames. Common methods include faster region-based convolutional neural network (faster R-CNN) [33]- [38], you only look once (YOLO) [39]- [43], single-shot multi-box detector (SSD) [44]- [47], and feature pyramid network (FPN) [48]- [51]. After training with a dataset, these methods recognize the external appearance of an object or animal of interest, and predict the blob position, size, and boundary of each animal.…”

Section: Introductionmentioning

confidence: 99%

Robust individual pig tracking

Jaoukaew,

Suwansantisuk,

Kumhom

2024

IJECE

View full text Add to dashboard Cite

The locations of pigs in the group housing enable activity monitoring and improve animal welfare. Vision-based methods for tracking individual pigs are noninvasive but have low tracking accuracy owing to long-term pig occlusion. In this study, we developed a vision-based method that accurately tracked individual pigs in group housing. We prepared and labeled datasets taken from an actual pig farm, trained a faster region-based convolutional neural network to recognize pigs’ bodies and heads, and tracked individual pigs across video frames. To quantify the tracking performance, we compared the proposed method with the global optimization (GO) method with the cost function and the simple online and real-time tracking (SORT) method on four additional test datasets that we prepared, labeled, and made publicly available. The predictive model detects pigs’ bodies accurately, with F1-scores of 0.75 to 1.00, on the four test datasets. The proposed method achieves the largest multi-object tracking accuracy (MOTA) values at 0.75, 0.98, and 1.00 for three test datasets. In the remaining dataset, the proposed method has the second-highest MOTA of 0.73. The proposed tracking method is robust to long-term occlusion, outperforms the competitive baselines in most datasets, and has practical utility in helping to track individual pigs accurately.

show abstract

Section: Introductionmentioning

confidence: 99%

Robust individual pig tracking

Jaoukaew,

Suwansantisuk,

Kumhom

2024

IJECE

View full text Add to dashboard Cite

show abstract

“…In recent times, the domain of computer vision has witnessed the substantial utilization of deep learning and convolutional neural network (CNN) methodologies. These advanced techniques have found extensive applications, especially in intricate assignments like image classification, semantic segmentation, and object detection [19,21]. With the constant improvement of computing resources and algorithm networks, deep learning models have demonstrated outstanding performance and reliable capability.…”

Section: Introductionmentioning

confidence: 99%

“…Currently, DL-based tree crown detection includes one-stage detection and two-stage detection [27]. One-stage networks typically output the position and category of each object in the image through a neural network [28], with the typical models including You Only Look Once (YOLO) [29] and Single Shot MultiBox Detector (SSD) [21]. They have the advantages of a fast detection speed and requiring few computational resources, making them suitable for real-time applications [30].…”

Section: Introductionmentioning

confidence: 99%

“…They have the advantages of a fast detection speed and requiring few computational resources, making them suitable for real-time applications [30]. For tree crown detection, Jintasuttisak et al [31] used the YOLOv5 network to achieve the fast and accurate detection of palm trees, while Zheng et al [21] used the SSD network for individual tree segmentation in Thai orchards, which greatly improved the accuracy compared to the traditional methods. Two-stage networks implement object detection in two stages.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Deep Learning Network for Individual Tree Segmentation in UAV Images with a Coupled CSPNet and Attention Mechanism

Lv,

Li,

Mao

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

Accurate individual tree detection by unmanned aerial vehicles (UAVs) is a critical technique for smart forest management and serves as the foundation for evaluating ecological functions. Existing object detection and segmentation methods, on the other hand, have reduced accuracy when detecting and segmenting individual trees in complicated urban forest landscapes, as well as poor mask segmentation quality. This study proposes a novel Mask-CSP-attention-coupled network (MCAN) based on the Mask R-CNN algorithm. MCAN uses the Cross Stage Partial Net (CSPNet) framework with the Sigmoid Linear Unit (SiLU) activation function in the backbone network to form a new Cross Stage Partial Residual Net (CSPResNet) and employs a convolutional block attention module (CBAM) mechanism to the feature pyramid network (FPN) for feature fusion and multiscale segmentation to further improve the feature extraction ability of the model, enhance its detail information detection ability, and improve its individual tree detection accuracy. In this study, aerial photography of the study area was conducted by UAVs, and the acquired images were used to produce a dataset for training and validation. The method was compared with the Mask Region-based Convolutional Neural Network (Mask R-CNN), Faster Region-based Convolutional Neural Network (Faster R-CNN), and You Only Look Once v5 (YOLOv5) on the test set. In addition, four scenes—namely, a dense forest distribution, building forest intersection, street trees, and active plaza vegetation—were set up, and the improved segmentation network was used to perform individual tree segmentation on these scenes to test the large-scale segmentation ability of the model. MCAN’s average precision (AP) value for individual tree identification is 92.40%, which is 3.7%, 3.84%, and 12.53% better than that of Mask R-CNN, Faster R-CNN, and YOLOv5, respectively. In comparison to Mask R-CNN, the segmentation AP value is 97.70%, an increase of 8.9%. The segmentation network’s precision for the four scenes in multi-scene segmentation ranges from 95.55% to 92.33%, showing that the proposed network performs high-precision segmentation in many contexts.

show abstract

“…The application of methods and data of GIS, remote sensing, and spatial analysis for forest and UGAs changes was carried out by Valjarević et al (2018), Jalkanen et al (2020), and Zheng and Wu (2021). In particular, Jalkanen et al (2020) used spatial prioritization methods to determine the importance of the UGAs.…”

Section: Introductionmentioning

confidence: 99%

Spatiotemporal Analysis of Urban Green Areas Using Change Detection: A Case Study of Kharkiv, Ukraine

Morar

Lukić

Valjarević

et al. 2022

Front. Environ. Sci.

View full text Add to dashboard Cite

The contemporary globalized world characterizes the rapid population growth, its significant concentration in cities, and an increase in the urban population. Currently, many socio-cultural, economic, environmental, and other challenges are arising in modern cities, negatively affecting the state of the urban environment, health, and quality of life. There is a need to study these problems in order to solve them. Urban Green Areas (UGAs) are a part of the social space and a vital part of the urban landscape. They act as an environmental framework of the territory and a factor ensuring a more comfortable environment of human life. This study aims at substantiating the importance of the UGAs, identifying the spatiotemporal dynamics of their functioning, and transforming changes in their infrastructure given the expansion of their functions. This research was carried out as a case study of the second city in Ukraine, Kharkiv. The authors developed and used an original integrated approach using urban remote sensing (URS) and GIS for changes detection to evaluate the current state and monitor spatial transformations of the UGAs. We used several GIS platforms and online resources to overcome the lack of digital cadastre of the thematic municipal area of Kharkiv. This resulted in the present original study. The study analyses the dynamics of the spatial and functional organization of the UGAs according to the Master Plans, plans, maps, and functional zoning of the city for the period from 1867 to 2019. The peripheral green areas became important after the large-scale development of the extensive residential areas during the rapid industrial development in remote districts of the city. They provide opportunities for population recreation near living places. Central UGAs are modern, comprehensively developed clusters with multidisciplinary infrastructure, while the peripheral UGAs are currently being developed. The use of URS/GIS tools in the analysis of the satellite images covering 2000–2020 allowed identifying the factors of the UGAs losses in Kharkiv and finding that UGAs were not expanding and partially shrinking during the study period. It is caused by the intensive construction of the residential neighborhoods, primarily peripheral areas, infrastructure development, and expansion of the city transport network. Nonetheless, some sustainable trends of UGA functioning without more or less significant decrease could be proved as existing in a long-term perspective. The authors analyzed and evaluated changes and expansion of the UGAs functions according to modern social demand. The research value of this is the usage of different approaches, scientific sources, URS/GIS tools to determine the UGAs transformation in the second-largest city in Ukraine (Kharkiv), to expand and update the main functions of UGAs and their role in the population’s recreation. The obtained scientific results can be used to update the following strategies, programs, and development plans of Kharkiv.

show abstract

Single Shot MultiBox Detector for Urban Plantation Single Tree Detection and Location With High-Resolution Remote Sensing Imagery

Cited by 8 publications

References 34 publications

Robust individual pig tracking

Robust individual pig tracking

A Deep Learning Network for Individual Tree Segmentation in UAV Images with a Coupled CSPNet and Attention Mechanism

Spatiotemporal Analysis of Urban Green Areas Using Change Detection: A Case Study of Kharkiv, Ukraine

Contact Info

Product

Resources

About